A hidden gene in astroviruses encodes a viroporin

Abstract

Human astroviruses are small non-enveloped viruses with positive-sense single-stranded RNA genomes. Astroviruses cause acute gastroenteritis in children worldwide and have been associated with encephalitis and meningitis in immunocompromised individuals. It is still unknown how astrovirus particles exit infected cells following replication. Through comparative genomic analysis and ribosome profiling we here identify and confirm the expression of a conserved alternative-frame ORF, encoding the protein XP. XP-knockout astroviruses are attenuated and pseudo-revert on passaging. Further investigation into the function of XP revealed plasma and trans Golgi network membrane-associated roles in virus assembly and/or release through a viroporin-like activity. XP-knockout replicons have only a minor replication defect, demonstrating the role of XP at late stages of infection. The discovery of XP advances our knowledge of these important human viruses and opens an additional direction of research into their life cycle and pathogenesis.

Fig. 1. Comparative genomic analysis of astroviruses.

a Phylogenetic tree of mammalian astroviruses. The tree, calculated with MrBayes, is based on ORF1b amino acid sequences obtained from 221 full-length genomes. Related groups of sequences (indicated by isosceles triangles) have been replaced in the figure by a single representative accession number and virus name; the total number of sequences in each group is shown in red (see Supplementary Fig. 1 for the complete tree). Genogroups (according to Yokoyama et al.) are indicated by background shading. Subgroups of sequences, defined for the purposes of this study only, are indicated in blue (Ia, Ib, etc). The tree is midpoint rooted and nodes are labeled with posterior probability values if different from 1.00. b Map of the human astrovirus genome showing the three main ORFs (blue) and the overlapping ORFX (pink), and analysis of conservation at ORF1a-ORF1b-ORF2 synonymous sites. The red line shows the probability that the observed conservation could occur under a null model of neutral evolution at synonymous sites, whereas the brown line depicts the ratio of the observed number of substitutions to the number expected under the null model. Inferred elements corresponding to regions of enhanced synonymous site conservation are indicated. c Analysis of an alignment of 127 human and 5 feline astrovirus ORF2 sequences. The upper three panels show the positions of alignment gaps (gray), stop codons (black), and AUG codons (green) in each reading frame. Below, is shown the analysis of conservation at synonymous sites. d Box plots of XP length for different astrovirus clades: centre lines = medians; boxes = interquartile ranges; whiskers extend to most extreme data point within 1.5 × interquartile range from the box; circles = outliers; number of sequences, n, is shown above each box. e Alignment of XP sequences from representative HAstVs. Source data are provided as a Source Data file.

Fig. 2. Ribosome profiling of astrovirus-infected cells.

Cells were harvested at 12 hpi and either flash frozen with no pre-treatment (NT), or pretreated with lactimidomycin for 30 min followed by flash-freezing (LTM). a RPF densities in reads per million mapped reads (RPM) for NT repeats, smoothed with a 15 nt sliding window. b RPF densities in RPM for LTM repeats, at single-nucleotide resolution. c For NT samples, phasing of 5′ ends of RPFs that map to host coding sequences, the part of ORF2 that is overlapped by ORFX, and the part of ORF2 that is not overlapped by ORFX. d Distributions of estimated XP-frame:CP-frame expression ratios for 100,000 bootstrap resamplings of codon positions within the overlapping and nonoverlapping regions of ORF2. Black lines show medians and 95% confidence intervals. Graphs show the results of two biologically independent experiments (#1 and #2). Source data are provided as a Source Data file.

Fig. 3. Design and properties of XP knockout viruses.

a Schematic representation of the astrovirus genome. Numbers correspond to astrovirus genome nucleotides in the pAVIC1 infectious clone; FS frameshift signal, SG subgenomic promoter. The sequence of nucleotides 4300–4400 is shown below, indicating the positions of overlapping elements and the corresponding translated proteins: ORF1b (purple), ORF2 (blue), and ORFX (red). b XP amino acid sequences for the wt and mutant pAVIC1-derived viruses († see main text for note on the AUGm mutant). See Supplementary Fig. 11 for nucleotide and CP amino acid sequences. c Titers (infectious units per mL, IU mL⁻¹) of recombinant viruses after RNA transfection of Huh7.5.1 cells followed by first passage (P1) and after seven passages (P7) in Caco2 cells. RT-PCR analysis of viral RNA isolated after P1 and P7 for selected recombinant viruses. See Supplementary Fig. 12 for sequencing chromatograms of individual passages.

Fig. 4. Analysis of replication stages affected by XP.

a Schematic of the astrovirus replicon. The 2A-RLuc cassette is fused in either the ORF2 (pO2RL) or ORFX (pXRL) reading frame. b List of replicon mutants showing the RLuc reading frame and introduced mutations. Asterisk—see Supplementary Fig. 11 for exact sequences. c, d Relative replicon luciferase activities measured after RNA transfection of BSR cells (see Supplementary Fig. 13a, b for Huh7.5.1 cells). Values are normalized so that the maximum ORF2-frame wt value is 100%. e Schematic of deletion mutants in the pO2RL replicon; ORF2 (blue), ORFX initiation codon (red). f Relative replicon luciferase activities of deletion mutants (see Supplementary Fig. 13c for Huh7.5.1 cells). g Schematic design and results of experiment to quantify virus titer, RNA, and protein levels in released virions and infected cells. BSR cells were electroporated with pAVIC1-wt, -AUGm or -PTC1 T7 RNAs and incubated for 48 h. Clarified supernatants were titrated (first graph) or treated with RNase I and used for viral RNA isolation and subsequent quantification by qRT-PCR (second graph). Cells were collected, freeze-thawed three times and titrated on Caco2 cells (third graph). Total RNA from infected cells was isolated and virus gRNA quantified by qRT-PCR (fourth graph). Aliquots of electroporated cells were seeded on a 96-well plate, incubated for 48 h, fixed, permeabilized, stained with anti-CP antibody, and imaged by LI-COR followed by quantification using LI-COR software (fifth graph). h Cells were collected at 24 hpe, freeze-thawed three times and titrated on Caco2 cells. i Aliquots corresponding to 10⁵ and 10⁴ electroporated cells were seeded on a 96-well plate, fixed at 24 hpe, permeabilized, stained with anti-CP antibody, and imaged by LI-COR followed by quantification using LI-COR software. j Huh7.5.1 cells were electroporated with pAVIC1-wt, -AUGm or -PTC1 T7 RNAs and incubated for 24 h. Representative confocal images of fixed and permeabilized cells visualized for CP (green) and stained for nuclei (Hoechst, blue). Images represent z-stack projections. Scale bars are 10 µm. Graphs show means ± s.d. from n = 3 biologically independent experiments (c, d, f, g, h). Source data are provided as a Source Data file.

Fig. 5. Cellular localization, membrane topology, and multimerization of XP.

a Huh7.5.1 cells were electroporated with pCAG-mCherry, pCAG-mCherry-XP, or pCAG-XP-mCherry. Representative confocal images of live cells stained for plasma membrane (WGA, green) and nuclei (Hoechst, blue); mCherry fluorescence is shown in red. HeLa cells were electroporated with pCAG-HA-XP, stained for plasma membrane (WGA, green), fixed, permeabilized and stained for XP with anti-HA antibody (red) and nuclei (Hoechst, blue). Images are averaged single plane scans. b HeLa cell lysates were fractionated and whole cell lysate (WCL), cytoplasmic (Cyto), membrane (Mem), and soluble nuclear (Nucl) fractions were analyzed by immunoblotting with antibodies to mCherry, tubulin, VDAC or Lamin A + C as indicated. See Supplementary Fig. 14 for complete images. c HeLa cells were electroporated with pCAG-XP-mCherry or pCAG-mCherry-XP. Cell surface mCherry on live cells was detected by incubation with anti-mCherry antibody, followed by staining with Alexa 488-labeled anti-rabbit IgG antibody and confocal microscopy. The images are averaged single plane scans. See Supplementary Fig. 15 for plasma membrane-permeabilized controls. A schematic representation of the observed membrane topologies is shown at right. d HeLa cells were electroporated with pCAG-XP-HA, fixed, permeabilized, and stained for XP (anti-HA, red), nuclei (Hoechst, blue), cis Golgi (anti-GM130, green), and trans Golgi (anti-TGN46, green). The images are averaged single plane scans. All scale bars are 10 µm (a,c,d). e Quantification of co-localization of XP-HA with TGN46 and GM130. The Pearson correlation coefficient was calculated for 12 images in each experiment. f Kyte-Doolittle hydropathy plots for HAstV XPs (see Supplementary Fig. 16 for individual plots). g C-terminal XP sequences for wt and 5L mutant HAstV1, and helical wheel representation of wt amino acids 92–109. h Huh7.5.1 cells were electroporated with wt or 5L mutant pCAG-Strep-XP or pCAG-HA-XP. At 16 hpe cells were lysed and subjected to immunoprecipitation using anti-HA magnetic beads. Presence of tagged proteins (5% of input and IP) was determined by western blotting using the indicated antibodies. See Supplementary Fig. 14 for complete images. Source data are provided as a Source Data file.

Fig. 6. XPs from different astroviruses have viroporin-like activity.

a Schematic representation of SINV replicon C used to evaluate cell-permeabilizing activity of expressed proteins. SINV elements: nsP1-4 – non-structural polyprotein; CP – capsid protein; SG – subgenomic promoter. b Membrane permeabilization in BSR cells at 8 h post RNA electroporation with Sindbis virus replicons (SINV repC) expressing HAstV1 XP, enterovirus Strep-2B (positive control), or mCherry (negative control). Ongoing protein synthesis was labeled with 1 mM AHA in the presence or absence of 1 mM HB as a translation inhibitor. Cells were lysed and AHA-bearing proteins were ligated to the fluorescent reporter IRDye800CW Alkyne by click chemistry, separated by SDS-PAGE, and visualized by in-gel fluorescence. Numbers below each pair of samples indicate protein synthesis quantified for HB-treated cells relative to values obtained for untreated cells. c Statistical analysis of membrane permeabilization caused by different XPs and mutants and the indicated control proteins in BSR cells. Bars indicate the amount of protein synthesis in HB-treated cells relative to untreated cells. P values are from comparisons with mock (BSR; black) or with pSINV-repC-XP (blue). d Sequences of wt, RR1, and RR2 HAstV1 mutants. e Sequences of HAstV1, HAstV4, and feline (FAstV), canine (CAstV) and porcine (PAstV) astrovirus XPs. f Alignment of pAVIC1-wt and -5L showing translation of XP (red) and the overlapping CP (blue). Introduced nucleotide and amino acid changes are highlighted in yellow. g Quantification of virus titer, RNA, and protein levels in released virions and infected cells as described in Fig. 4g. h Caco2 cells were infected with the indicated viruses at MOI 0.2 and incubated for 48 h. Released virus in clarified supernatants was titrated. i Caco2 cells were infected with the indicated viruses at MOI 0.2 in the presence (dark blue) or absence (light blue) of 5 µM hexamethylene amiloride (HMA). Intracellular (HMA-free) virus was titrated. See Supplementary Fig. 20 for cell toxicity data. P values come from two-tailed t-tests without adjustment for multiple comparisons. Error bars indicate mean ± s.d.; n = 3 (c, g, h) or 4 (i) biologically independent experiments. Source data are provided as a Source Data file.